10.24425/acs.2022.

142848
Archives of Control Sciences
Volume 32(LXVIII), 2022
No. 3, pages 557–576

Finite-time SDRE control of F16 aircraft dynamics

Marcin CHODNICKI, Paweł PIETRUSZEWSKI, Mariusz WESOŁOWSKI and Sławomir STĘPIEŃ

This paper proposes a finite-time horizon suboptimal control strategy based on state-
dependent Riccati equation (SDRE) to control of F16 multirole aircraft. Flight stabilizer control
of super maneuverable aircraft is modelled and simulated. For aircraft modelling purpose a
full 6 DOF model is considered and described by nonlinear state-space approach. Also a stable
state-dependent parametrization (SDP) necessary for solution of the SDRE control problem is
proposed. Solution of the SDRE control problem with adequate defined weighting matrices
in performance index shows possibility of fast and optimal aircraft control in finite-time. The
method in this form can be used for stabilization of aircraft flight and aerodynamics.
Key words: aircraft modelling, state-dependent Riccati equations, finite-time optimal con-
trol

1. Introduction

Flight control of multirole fighter aircraft is viewed as a difficult and problem-

atic area of aerospace engineering, primarily since it has a track record associated
with program delays, and aircraft incidents and accidents [34–36]. Modern and
innovative aircraft are equipped with robust and optimal controllers, LQG/LTR
in longitudinal control for instance. But, the primary purpose of a flight control
system is to provide the appropriate interface between a pilot and the aircraft re-
sponses. Although stabilization is a requirement of a typical modern FCS (Flight
Control Systems), it is stabilization with a pilot in the loop, or flying qualities,
Copyright © 2022. The Author(s). This is an open-access article distributed under the terms of the Creative Com-
mons Attribution-NonCommercial-NoDerivatives License (CC BY-NC-ND 4.0 https://creativecommons.org/licenses/
by-nc-nd/4.0/), which permits use, distribution, and reproduction in any medium, provided that the article is properly
cited, the use is non-commercial, and no modifications or adaptations are made
M. Chodnicki, P. Pietruszewski and M. Wesołowski are with Air Force Institute of Technology, Księcia
Bolesława 6, 01-494 Warsaw, Poland.
S. Stępień (corresponding author, e-mail: slawomir.stepien@put.poznan.pl) is with Poznan University
of Technology, Institute of Automatic Control and Robotics, Piotrowo 3a, 60-965 Poznań, Poland.
We would like to express our sincere appreciation to flight instructor and aerobatic pilot Tadeusz
Zaworski from Air Force Institute of Technology in Warsaw and Lt.-Col. pilot Adam Wójcik from Krzesiny
NATO Air Base for their valuable and constructive talks, suggestions and comments during writing and
development of this research work.
Received 17.01.2022. Revised 11.08.2022.
558 M. CHODNICKI, P. PIETRUSZEWSKI, M. WESOŁOWSKI, S. STĘPIEŃ

that is the design challenge and has caused most problems [1, 29, 36]. The re-
lationship between pilot action and aircraft reaction should be fast, performed
in finite time, eliminating aerodynamics nonlinearities, uncertainties or uncon-
trollable oscillations resulting from efforts of the pilot to control the aircraft and
occur when the pilot of an aircraft inadvertently commands an often increas-
ing series of corrections in opposite directions, known as PIO (Pilot Involved
Oscillations) [19, 36].
One of popular and very effective is a single-engine multirole fighter aircraft
F16 originally developed by General Dynamics for the United States Air Force
(USAF) also used by Polish Air Force (PAF). The multirole F16 aircraft is shown
in Fig. 1.

Figure 1: F16 multirole fighter aircraft [37]

Many papers deal with the F16 aircraft dynamics, where finite-time opti-
mal control methods used to flight control systems is still a challenge for many
engineers and researchers [1,18,22,23,29,34,35]. Literature provide some inter-
esting works related to finite-time problems. In [12] authors studied a finite-time
sliding mode attitude controller for a reentry vehicle with blended aerodynamic
surfaces and a reaction control system. Next work [29] investigated the path
planning of a reusable launch vehicle using a finite-horizon suboptimal con-
troller. Further, the paper [9] deals with digital controller for longitudinal aircraft
model, where the control task is formulated as a tracking problem of velocity and
flight path angle considering incomplete information about varying parameters
of the system and external disturbances. The output tracking control, employ-
ing model reference feedback linearization of an aircraft subject to additive,
uncertain, nonlinear disturbances is presented and described in [27]. An inter-
esting control problem is studied in [31] where the finite-time attitude tracking
 FINITE-TIME SDRE CONTROL OF F16 AIRCRAFT DYNAMICS 559

control problem of a reusable launch vehicle in the reentry phase under in-
put constraints using a constrained adaptive back-stepping fast terminal sliding
mode control technique. A nonsingular terminal sliding-mode control method
with finite-time fault-tolerant control for spacecraft with actuator saturations is
studied in [18].
Nowadays, modern optimal control theory proposes high performance and
rapidly emerging control technique called finite-time SDRE (State-Dependent
Riccati Equation) [2, 7, 20, 24, 32]. This is a suboptimal control methodology for
nonlinear systems. The technique uses direct parameterization to bring the nonlin-
ear system to a linear structure having state-dependent coefficients (SDC) [33].
A state-dependent Riccati equation (SDRE) is then solved accordingly to the
change of state trajectory to obtain a nonlinear feedback controller matrix, which
coefficients, in other feedback gains are the solution of SDRE [8].
The method, firstly proposed in 1962 [21] and later expanded in 1975 [32],
was further carefully analyzed and deeply studied creating its useful form for
technical applications [20]. The method employs parameterization of the nonlin-
ear dynamics into the state vector, then the product of a matrix-valued function
depends on the state itself [33]. The control technique fully captures the non-
linearities of the dynamic system, bringing the system to a (nonunique) linear
structure having state-dependent coefficient (SDC) matrix form, and minimiz-
ing also a state-dependent nonlinear performance index having a quadratic-like
structure. The differential SDRE equations using the SDC matrices is then solved
on-line (in real time) to give the suboptimum control law. The technique for
the finite-time nonlinear optimal control problem in the multivariable case is
locally asymptotically stable and locally asymptotically optimal as described in
following theoretical contributions [2, 4, 6, 16, 20].
Applications of the SDRE control technique include also satellite and space-
craft control and estimation, integrated guidance and control design, autopilot
design, robotics, control of systems with parasitic effects, control of artificial
human pancreas, ducted fan control and magnetic systems including levitation
and drives [2, 7, 18].
In this paper, a modelling and control design methodology as the concept
is proposed to design of high-performance and optimal flight stabilizer for F16
multitask military aircraft. The paper presents a nonlinear model of the aircraft
and solution of the finite-time suboptimal control problem for flight stabilization
problem minimizing energy lost and delivered to the flaying machine.

2. F16 nonlinear model

The rigid body equations of motion are the differential equations that describe
the evolution of basic states of an aircraft. The aircraft model presents Fig. 2.
560 M. CHODNICKI, P. PIETRUSZEWSKI, M. WESOŁOWSKI, S. STĘPIEŃ

Figure 2: F16 aircraft model

The aircraft dynamics is generally defined using Newton’s force and moment
equations [34, 35]. The force equation is following

F = 𝑚 (¤v + 𝝎 × v) , (1)

where v is an aircraft linear velocity vector, 𝝎 is angular velocity vector, 𝑚 is

an aircraft mass and F denotes, of course, force vector. For completeness, also
moment equation should be considered. The equation describes all the moments
acting on the aircraft, equal to the rate of change of angular momentum

¤ + 𝝎 × I𝝎,
M = I𝝎 (2)

where I is an aircraft inertia matrix and M denotes moment vector. When consider
vector v defined for all components in 𝑥, 𝑦 and 𝑧 direction and 𝝎 for roll 𝜑, pitch
𝜃 and yaw 𝜓 angle
" # " #
𝑢 𝑝
v= 𝑣 and 𝝎 = 𝑞 (3)
𝑤 𝑟
then equations of aircraft aerodynamics can be defined for linear and angular
speeds. In addition, because of a plane of symmetry so in the inertia matrix the
cross-products involving 𝑦 become zero

𝐼𝑥𝑥 0 −𝐼𝑥𝑧
" #
I = 0 𝐼 𝑦𝑦 0 . (4)
−𝐼 𝑧𝑥 0 𝐼 𝑧𝑧
 FINITE-TIME SDRE CONTROL OF F16 AIRCRAFT DYNAMICS 561

The system of nonlinear equations that describes aircraft flight dynamics, con-
sidering gravity forces 𝑔 and force due to the thrust 𝐹𝑇 , is following

" # 𝑟𝑣 − 𝑞𝑤 + 𝑔 sin 𝜃 + 1 𝐹 

𝑢¤ 𝑇
 𝑚 
𝑣¤ =  𝑝𝑤 − 𝑟𝑢 − 𝑔 sin Ø cos 𝜃  , (5)
𝑤¤ 
 𝑞𝑢 − 𝑝𝑣 − 𝑔 cos Ø cos 𝜃 

 
 𝑝¤ 
   𝑐 1 𝑝𝑞 + 𝑐 2𝑟𝑞 + 𝑐 𝑙 𝑝 𝐿 + 𝑐 𝑛𝑝 𝑁 

 𝑞¤  = 𝑐 3 𝑝 2 − 𝑟 2
+ 𝑐 4 𝑝𝑟 + 𝑐 𝑚𝑞 (𝑀 + 𝐹𝑇 𝑍𝑇 𝑃 )  , (6)
   
 𝑟¤   𝑐 5 𝑝𝑞 + 𝑐 6𝑟𝑞 + 𝑐 𝑙𝑟 𝐿 + 𝑐 𝑛𝑟 𝑁 
   
 𝑇
where 𝐿, 𝑀, 𝑁 are components of the aircraft moment vector M = 𝐿 𝑀 𝑁

2

𝑐 1 = −𝐼 𝑧𝑧 𝐼𝑥𝑧 − 𝐼𝑥𝑧 (𝐼𝑥𝑥 − 𝐼 𝑦𝑦 ) / 𝐼𝑥𝑧 − 𝐼𝑥𝑥 𝐼 𝑧𝑧 ,

2
2

𝑐 2 = − 𝐼 𝑧𝑧 𝐼 𝑦𝑦 − 𝐼 𝑧𝑧 + 𝐼𝑥𝑧 / 𝐼𝑥𝑧 − 𝐼𝑥𝑥 𝐼 𝑧𝑧 ,
𝑐 3 = −𝐼𝑥𝑧 /𝐼 𝑦𝑦 , 𝑐 4 = − (𝐼𝑥𝑥 − 𝐼 𝑧𝑧 ) /𝐼 𝑦𝑦 ,
2

2

𝑐 5 = − 𝐼𝑥𝑧 − 𝐼𝑥𝑥 𝐼𝑥𝑥 − 𝐼 𝑦𝑦 / 𝐼𝑥𝑧 − 𝐼𝑥𝑥 𝐼 𝑧𝑧 ,


2

𝑐 6 = 𝐼𝑥𝑥 𝐼𝑥𝑧 − 𝐼𝑥𝑧 𝐼 𝑦𝑦 − 𝐼 𝑧𝑧 / 𝐼𝑥𝑧 − 𝐼𝑥𝑥 𝐼 𝑧𝑧 ,
2

𝑐 𝑙 𝑝 = 𝑐 𝑛𝑝 = −𝐼 𝑧𝑧 / 𝐼𝑥𝑧 − 𝐼𝑥𝑥 𝐼 𝑧𝑧 ,
𝑐 𝑚𝑞 = 1/𝐼 𝑦𝑦 ,
2
2

𝑐 𝑙𝑟 = −𝐼𝑥𝑧 / 𝐼𝑥𝑧 − 𝐼𝑥𝑥 𝐼 𝑧𝑧 , 𝑐 𝑛𝑟 = −𝐼𝑥𝑥 / 𝐼𝑥𝑧 − 𝐼𝑥𝑥 𝐼 𝑧𝑧
and 𝑍𝑇 𝑃 is a position of engine thrust point.
Equations (5)–(6) are a nonlinear vector and it has to be formed as SDC ma-
trices. Separation of (5)–(6) is not so complicated for aircraft systems, because
in general, state variables are in the form of products and it make parametriza-
tion easy. Only terms related to gravity 𝑔 seems to be problematic in sense of
parametrization. But for the case study purpose, for example, when consider air-
craft flight with nonzero speed, the terms can be parametrized by the speed in
flight direction.
To describe the aircraft orientation an Euler angle relationship, in other Euler
angles for flat Earth assumption, is used from the transformation from the local
horizontal to the body axes. The resulting kinetic equations are
 𝜙¤   𝑝 + (𝑞 sin 𝜙 + 𝑟 cos 𝜙) tan 𝜃 
 𝜃¤  = 
   
   𝑞 cos 𝜙 − 𝑟 sin 𝜙 ,
 (7)
𝜓¤   (𝑞 sin 𝜙 + 𝑟 cos 𝜙) sec 𝜃 
   
where 𝜙 is a roll angle, 𝜃 is a pitch angle, and 𝜓 is a yaw angle and sec 𝜃 = 1/cos 𝜃.
562 M. CHODNICKI, P. PIETRUSZEWSKI, M. WESOŁOWSKI, S. STĘPIEŃ

3. SDRE control problem

The optimal control method is well described in [2, 5, 7, 8, 24]. Interested

scientists and readers can follow the state-dependent Riccati equation (SDRE)
approach in the context of the nonlinear regulator problem with quadratic objec-
tive function [11, 13, 14, 20].
The finite-time control problem consists of finding optimal control law that
minimizes following objective function defined for final control time 𝑡 𝑓 [32]
∫𝑡 𝑓  
1 1
𝐽 (u) = x𝑇 (𝑡 𝑓 )Sx(𝑡 𝑓 ) + 𝑇 𝑇
x Q(x)x + u R(x)u d𝑡, (8)
2 2
0
subject to nonlinear dynamics for affine systems
x¤ = F(x) + B(x)u. (9)
Nonlinear dynamics (9) can be written using the state-dependent coefficient
(SDC) form [17, 33]
x¤ = A(x)x + B(x)u. (10)
where S(x) and Q(x) are symmetric, positive semi-definite weighting matrices
for states, R(x) is the symmetric, positive definite weighting matrix for control
inputs. Equation (9) includes F(x) vector, which is piecewise continuous in time
and smooth respect to their arguments, which satisfy the Lipschitz condition.
Considering (10), if the pair {A(x), B(x)} is a stabilizable parameterization
of the system, then to check controllability of the affine system, this pair in linear
sense should be controllable. In other, checking the controllability of that pair
does not need the state or control input information [13]. It can be simply checked
by the matrix
 
M(x) = B(x) A(x)B(x) . . . A𝑛−1 (x)B(x) (11)
often called controllability matrix. Then the system (9) or (10) is controllable if
the controllability matrix (11) has full rank.
Employing the Hamiltonian theory
1 𝑇 
H= x Q(x)x + u𝑇 R(x)u + p𝑇 (A(x)x + B(x)u) (12)
2
𝜕H
and considering the necessary optimality condition = 0 with p = K(x)x,
𝜕u
results in a control law as
u = −-R(x) −1 B(x)𝑇 K(x)x. (13)
The control law (13) includes state-dependent feedback compensator K(x) which
is a solution of SDRE.
 FINITE-TIME SDRE CONTROL OF F16 AIRCRAFT DYNAMICS 563

Considering optimality conditions [2, 7, 20]:

𝜕H 𝜕H
p¤ = − and x¤ = − (14)
𝜕x 𝜕p
the nonlinear system is described by state-space equation

x¤ = A(x)x − B(x)R(x) −1 B(x)𝑇 p (15)

and following adjoint differential equation

 𝑇  𝑇
𝜕 (A(x)x) 𝜕B(x)
p¤ = − p− u p − Q(x)x
𝜕x 𝜕x
1 𝜕Q(x) 1 𝜕R(x)
− x𝑇 x − u𝑇 u, (16)
2 𝜕x 2 𝜕x
where
𝜕 (A(x)x) 𝜕A(x)
= A (x) + x (17)
𝜕x 𝜕x
Substituting p = K(x)x into (16), next mathematical operations provide the
following differential equation
¤
K(x)x + K(x)A(x)x + A(x)𝑇 K(x)x − K(x)B(x)R(x) −1 B(x)𝑇 K(x)x
 𝑇  𝑇
𝑇 𝜕A(x) 𝑇 𝜕B(x)
+ Q(x)x + x K(x)x + u K(x)x
𝜕x 𝜕x
1 𝜕Q(x) 1 𝜕R(x)
− x𝑇 x − u𝑇 u = 0. (18)
2 𝜕x 2 𝜕x
Dimensions of derivatives of matrices B(x), A(x), Q(x) and R(x) respect to x,
do not have the same dimension of Eq. (18), so it must be rewritten as
¤
K(x) + K(x)A(x) + A(x)𝑇 K(x) − K(x)B(x)R(x) −1 B(x)𝑇 K(x) + Q(x)
 𝑇  𝑇
𝜕A(x) 𝜕B(x)
+ x K(x) + u K(x)
𝜕x 𝜕x
 𝑇  𝑇
1 𝜕Q(x) 1 𝜕R(x)
+ x − u R(x) −1 B(x)𝑇 K(x) = 0 (19)
2 𝜕x 2 𝜕x
Equation (19) is a nonlinear differential optimal control equation which is the
result of applying the optimality conditions and feedback control on Hamilton-
Jacobi-Bellman equation (HJB). The solution to equation (19) obtains K(x) which
564 M. CHODNICKI, P. PIETRUSZEWSKI, M. WESOŁOWSKI, S. STĘPIEŃ

is the nonlinear optimal gain. In order to extract the SDRE-like form of equation
(19), it is assumed that K(x) is the suboptimal solution to
¤
K(x) + K(x)A(x) + A(x)𝑇 K(x)
− K(x)B(x)R(x) −1 B(x)𝑇 K(x) + Q(x) = 0 (20)

with boundary condition K(x 𝑓 ) = S(x(𝑡 𝑓 )), then

 𝑇  𝑇
𝜕A(x) 𝜕B(x)
x K(x) + u K(x)
𝜕x 𝜕x
 𝑇  𝑇
1 𝜕Q(x) 1 𝜕R(x)
+ x − u R(x) −1 B(x)𝑇 K(x) = 0 (21)
2 𝜕x 2 𝜕x
is assumed and known as the optimality condition.
Equation (20) is in the form of differential SDRE for affine systems. Solution
of the equation results in suboptimal control because it neglects (21), which is
so-called “SDRE necessary condition for optimality” and it tends to zero [14].
Equation (20) known as differential SDRE or shortly SDDRE (State-
Dependent Differential Riccati Equation), can be solved numerically employing
different algorithms [15]. Today, very popular and classic is backward integra-
tion method to solve optimal control problem with a final boundary condition.
This approach is easy to implement but unfortunately has one disadvantage, it
provides two-rounded solution. The first solution generates optimal gain and the
second solution gives the complete answer, called two-rounded backward and
forward solutions. Another approach is based on state transition matrix technique
to solve SDRE in finite-time horizon. In this approach, the starting conditions
deal with introducing the state and co-state vectors obtained from Hamiltonian.
Then considering the boundary condition type free, the two-level solution of the
problem is reduced to one. This advantage makes the algorithm more optimal.
The value of final time should be reasonable and adequate to the system dynam-
ics. It is necessary to avoid numerical difficulties due to the inverse of the near
singular transition matrix, which provides the optimal matrix gain. In general,
to big values of the final time may produce computational difficulty where leads
to generate a singularity problem. Finally, another commonly used technique
for solving finite-time optimal control problem related to differential SDRE is
Lyapunov based method. Using this approach, the feedback control gain can be
computed directly. The method allows to get the positive-definite suboptimal
matrix gain. It means that this procedure can work with both steady-state roots
of SDRE. Then of course, the proofs of positive definiteness of the gain in both
cases: for largest and smallest solution are necessary. The magic of this approach
is that it works with the largest and smallest solutions, by substituting in resulting
 FINITE-TIME SDRE CONTROL OF F16 AIRCRAFT DYNAMICS 565

optimal gain. However, the numerical difficulties when both solutions are used
become problematic in this technique, due to computation of co-state vector.
The vector sometimes may take big values and the inverse of it, may be close to
singularity.

4. Stability proof

Asymptotic stability of the closed-loop system

x¤ = A(x)x − B(x)R(x) −1 B(x)𝑇 K(x)x (22)

implies that it is possible to control the states from the initial values to the final
ones. The stability the SDDRE can be presented via Lyapunov approach. The
Lyapunov candidate function is structured as

V(x) = x𝑇 K(x)x. (23)

Taking the time derivative of Lyapunov candidate function and substituting with
equations (13) and (20), results in
 
¤
V(x) = −x𝑇 K(x)B(x)R(x) −1 B(x)𝑇 K(x) + Q(x) x (24)

which is non-negative since the matrices Q(x), R(x) and K(x) are positive-definite
and B(x)R(x) −1 B(x)𝑇 is also non-negative.
In summary, having derivative of Lyapunov function (23), the local exponen-
tially stability, and hence the local uniform asymptotic stability of the system (22)
follows [2, 7, 20]. The local stability result cannot be readily generalized to the
global stability, however, such generalization can be carried out through utilizing
the concept of region of attraction [14] for defined bounded set
¤
  
𝛀 = x ∈ R𝑛 , V(x) ¬ 𝑐, 𝑡 ∈ 𝑡 0 , 𝑡 𝑓 , (25)
where 𝑐 is a positive constant.

5. Control system analysis

The nonlinear F16 aircraft model is applied to check the described finite-
time SDRE control for flight stabilization when the wind or other external forces
try to destabilize aircraft fligt-path. Governing equations that describe aircraft
aerodynamics are given by (5)-(6), but for the control purpose, state-dependent
parametrization SDC is necessary. When considering the flight dynamics for
566 M. CHODNICKI, P. PIETRUSZEWSKI, M. WESOŁOWSKI, S. STĘPIEŃ

nonzero speed 𝑢 ≠ 0, parametrized F16 model (10) based on system (5) and (6)
with gravity compensation, can be described in SDC form

 𝑢¤   (𝑔 sin 𝜃)/𝑢 0 0 0 −𝑤 𝑣 
 𝑢 
(−𝑔 sin Ø cos 𝜃)/𝑢 −𝑢 
    
 𝑣¤   0 0 𝑤 0 𝑣 
𝑤¤   (−𝑔 cos Ø cos 𝜃)/𝑢 0 −𝑣
    
0 𝑢 0  𝑤 
 𝑝¤  =  
0 0 𝑐 1 𝑝 + 𝑐 2𝑟
 𝑝
   0 0 0 
  
 𝑞¤  
𝑐 4 𝑝 − 𝑐 3𝑟 
𝑞
   0 0 0 𝑐3 𝑝 0  
 𝑟¤  
0 0 𝑐 5 𝑝 + 𝑐 6𝑟
𝑟 
   0 0 0 
  
 1/𝑚 0 0 0 

 0 0 0 0  𝐹𝑇 

 0 0 0 0   𝐿 
+  , (26)
 0 𝑐 𝑙𝑝 0 𝑐 𝑛𝑝   𝑀 
𝑐 𝑚𝑞 𝑍𝑇 𝑃 0 𝑐 𝑚𝑞 0   𝑁 

 0 𝑐 𝑙𝑟 0 𝑐 𝑛𝑟 

where control vector consists of thrust 𝐹𝑇 , and rolling, pitching, yawing moments
resulted from arrangements of ailerons, elevators and rudder [34].
As shown in Fig. 2, the thrust acts positively along the positive body 𝑥-axis.
Positive thrust cause an increase in acceleration along the body 𝑥-axis. For the
other control surfaces a positive deflection gives a decrease in the body rates.
A positive aileron deflection 𝛿𝑎 gives a decrease in the roll rates, this requires that
the right aileron deflect downward and the left aileron deflect upward. A positive
elevator deflection 𝛿 𝑒 results in a decrease in pitch rate, thus elevator is deflected
downwards. Positive deflection of the rudder 𝛿𝑟 decreases the yaw rate, and can
be described as a deflection to right. The maximum control values and units are
listed in Table 1.
Table 1: Minimum and maximum control values

Control Value [quantity]

𝐹𝑇 ∼ 150 [kN]
𝛿𝑒 −25÷25 [deg]
𝛿𝑎 −21.5÷21.5 [deg]
𝛿𝑟 −30÷30 [deg]

Considering defined real controls, arranged by aircraft pilot, neglecting lead-

ing edge flaps effect and assuming quasi-linear relationship between aircraft
moments and above described controls (the linearization is possible at the trim
point), the state-space system of equation (26) can be recalculated for new controls
 FINITE-TIME SDRE CONTROL OF F16 AIRCRAFT DYNAMICS 567

presented in Table 1

 𝑢¤   (𝑔 sin 𝜃)/𝑢 0 0 0 −𝑤 𝑣  𝑢 
 
 𝑣¤   (−𝑔 sin Ø cos 𝜃)/𝑢 −𝑢   𝑣 
  
0 0 𝑤 0
𝑤¤   (−𝑔 cos Ø cos 𝜃)/𝑢 0 −𝑣
    𝑤 
0 𝑢 0
 𝑝¤  =   
   0 0 0 0 𝑐 1 𝑝 + 𝑐 2𝑟 0   𝑝
 
 𝑞¤  
   0 0 0 𝑐3 𝑝 0 𝑐 4 𝑝 − 𝑐 3𝑟   𝑞 
 𝑟¤  
   0 0 0 0 𝑐 5 𝑝 + 𝑐 6𝑟 0   𝑟 

 1/𝑚 0 0 0 
 
 0 0 0 0  𝐹𝑇 
  
 0 0 0 0   𝛿𝑒 
+    , (27)
 0 0 𝑐 𝑙 𝑝 𝑐 𝑙1 + 𝑐 𝑛𝑝 𝑐 𝑛1 𝑐 𝑙 𝑝 𝑐 𝑙2 + 𝑐 𝑛𝑝 𝑐 𝑛2   𝛿𝑎 
𝑐 𝑚𝑞 𝑍𝑇 𝑃 𝑐 𝑚𝑞 𝑐 𝑚1
 0 0   𝛿𝑟 
 
 0
 0 𝑐 𝑙𝑟 𝑐 𝑙1 + 𝑐 𝑛𝑟 𝑐 𝑛1 𝑐 𝑙𝑟 𝑐 𝑙2 + 𝑐 𝑛𝑟 𝑐 𝑛2 

where coefficients 𝑐 𝑙1 , 𝑐 𝑙2 , 𝑐 𝑚1 , 𝑐 𝑛1 , 𝑐 𝑛2 are derived from the density of air flowing

in and Mach number, the geometry of the wing structure including its span, the
mean aerodynamic chord and wing area.
The F16 aircraft properties [1,23] used with certain assumptions and indicated
values to be able to perform further calculations in the chapter due to the model
(5)–(6) and (26)–(27) are presented in Table 2.
Table 2: Mass, inertia and other properties of F16

Parameter Value [quantity]

𝑚 9295 [kg]
𝐼𝑥 𝑥 12874.8 [kg × m2 ]
𝐼𝑦 𝑦 75673.6 [kg × m2 ]
𝐼 𝑧𝑧 85552.1 [kg × m2 ]
𝐼 𝑥𝑧 1331.4 [kg × m2 ]
𝑍𝑇 𝑃 0.3 [m]
𝑐 𝑙1 0.8
𝑐 𝑙2 0.2
𝑐 𝑚1 1.2
𝑐 𝑛1 0.2
𝑐 𝑛2 0.8

Employing described aircraft model, the SDRE control method is applied to

control the flight stabilization problem, considering two values of finite control
times: 𝑡 𝑓 = 3 s and 𝑡 𝑓 = 5 s. As mentioned in introduction, the time is very
568 M. CHODNICKI, P. PIETRUSZEWSKI, M. WESOŁOWSKI, S. STĘPIEŃ

important, because the aircraft should rapidly answer for pilot commands. The
path of flight must be sometimes rapidly stabilized when unexpected external
forces try to change the aircraft position and orientation during flying action.
Considering above, the problem consists of finding F16 aircraft state dynamics
and SDRE controls. In association with the aircraft dynamics (26), the quadratic
cost functional weighting matrices in (8) are chosen as
 1 0.15 0 0 0 0
 
0.15 1 0 0 0 0
 
3  0 0 1 0 0 0
S = 5 · 10  , Q = 5 · 103 I6×6 .
 0 0 0 1 0 0
 0 0 0 0 1 0

 0 0 0 0 0 1

and R = 2 · 10−4 I4×4 with initial speeds 𝑢 = 300 km/h, 𝑣 = −20 km/h, 𝑤 =
20 km/h and angle 𝜃 = 5◦ , in other hand, presenting it in vector form x0 =
 𝑇
300 · 0.28 − 25 · 0.28 20 · 0.28 0 5 · 𝜋/180 0 , where 0.28 is a constant that
allows to recalculate the speed in km/h to m/s and 𝜋/180 allows to recalculate
degs to rads. The control should stabilize the aircraft in finite time 𝑡 𝑓 and uphold
linear speed 𝑢 on prescribed level, it means that the final state conditions should
 𝑇
be as follows, for instance x 𝑓 = 500 · 0.28 0 0 0 0 0 . Considering the
final state, control law (13) takes the form
u = −R(x) −1 B(x)𝑇 K(x)(x − x 𝑓 ). (28)
Simulations are done to show the performance of the control designed in Section 3.
The aircraft state dynamics, controls and flight trajectory from initial state to the
final state are shown below.
Firstly, simulations are performed for the final time 𝑡 𝑓 = 3 s.
Next simulations are performed for the final time 𝑡 𝑓 = 5 s.
Figures 3–12 show closed-loop response of the flight controller and F16
aircraft system. Simulations are performed for two assumptions of the final control
time 𝑡 𝑓 : 3 and 5 seconds. Simulation time for both cases is two times longer than
control time 𝑡 𝑓 , because it is interesting haw the system works after 𝑡 𝑓 when
transversality condition holds.
When look at Figs. 3 and 8, it is worth to observe that for 𝑡 𝑓 = 3 s, the aircraft
is faster stabilized by feedback control and increase altitude 𝑧 only within 3.1
meters and distance 𝑥 approximately 202 meters. For the 𝑡 𝑓 = 5 s, the aircraft
is also stabilized well, but the aircraft change the altitude to 4.6 meters at the
distance 𝑥 equal to 585 meters. Hence the aircraft can be successively controlled
within assumed finite control times.
When consider controls, firstly thrust and aircraft moments shown at Figs. 4a–
4b and Figs. 9a–9b, further elevators, ailerons and rudder functions of deflection,
FINITE-TIME SDRE CONTROL OF F16 AIRCRAFT DYNAMICS 569

Figure 3: Aircraft flight trajectory

(a) Thrust force control

(b) Rolling, pitching and yawing moments

Figure 4
570 M. CHODNICKI, P. PIETRUSZEWSKI, M. WESOŁOWSKI, S. STĘPIEŃ

Figure 5: Elevators, ailerons and rudder deflection

Figure 6: Aircraft speeds

Figure 7: Roll, pitch and yaw angles

FINITE-TIME SDRE CONTROL OF F16 AIRCRAFT DYNAMICS 571

Figure 8: Aircraft flight trajectory

(a) Thrust force control

(b) Rolling, pitching and yawing moments

Figure 9
572 M. CHODNICKI, P. PIETRUSZEWSKI, M. WESOŁOWSKI, S. STĘPIEŃ

Figure 10: Elevators, ailerons and rudder deflection

Figure 11: Aircraft speeds

Figure 12: Roll, pitch and yaw angles

 FINITE-TIME SDRE CONTROL OF F16 AIRCRAFT DYNAMICS 573

the maximum thrust is kept longer for shorted control time 𝑡 𝑓 = 3 s. Moments
and deflection functions are similar for both cased, except yawing moments and
rudder deflections. The thrust is kept longer for 𝑡 𝑓 = 3 s because the aircraft must
faster stabilize and straighten flight trajectory.
Analysis of the aircraft speeds, presented at Figs. 6 and 11, proof that speed
in the flight direction is increased to 500 km/h with simultaneously zeroing other
speeds for both cases. The situation is consistent to assumed starting and final
state conditions. Additionally, the solution of kinematics is presented at Figs. 7
and 12. There is shown how evaluates the aircraft space orientation.

6. Conclusions

The finite time control problem for flight stabilization in flight control system
in multirole F16 aircraft is formulated and solved. The nonlinear, state-dependent
parametrized model of the aircraft is proposed. The optimal control technique
with nonlinear feedback compensator for computation of the control input that
minimizes energy delivered to the aircraft system and energy lost, performing
stabilization task is analyzed. The effectiveness of presented technique is demon-
strated on a numerical example where optimal aircraft controls are found for
different final times. Presented results proof, that proposed SDRE technique can
be successively applied to the F16 flight control aircraft systems.

References

[1] W. Ahmed, Z. Li, H. Maqsood, and B. Anwar: System modeling and

controller design for lateral and longitudinal motion of F-16. Automation,
Control and Intelligent Systems, 7(1), (2019), 39–45. DOI: 10.11648/j.acis.
20190701.15.

[2] H.T. Banks, B.M. Lewis, and H.T. Tran: Nonlinear feedback controllers
and compensators: a state-dependent Riccati equation approach. Com-
putational Optimization and Applications, 37(2), (2007), 177–218. DOI:
10.1007/s10589-007-9015-2.

[3] J.C.A. Barata and M.S. Hussein: The Moore-Penrose Pseudoinverse. A

Tutorial Review of the Theory. John Hopkins University Press, 2013.

[4] A. Bracci, M. Innocenti, and L. Pollini: Estimation of the region of attrac-

tion for state-dependent Riccati equation controllers. Journal of Guidance,
Control and Dynamics, 29(6), (2006), 1427–1430. DOI: 10.2514/1.22122.
574 M. CHODNICKI, P. PIETRUSZEWSKI, M. WESOŁOWSKI, S. STĘPIEŃ

[5] T. Cimen and S.P. Banks: Global optimal feedback control for general non-
linear systems with non-quadratic performance criteria. System and Control
Letters, 53(5), (2004), 327–346. DOI: 10.1016/j.sysconle.2004.05.008.
[6] T. Cimen and S.P. Banks: Nonlinear optimal tracking control with appli-
cation to super-tankers for autopilot design. Automatica, 40(11), (2004),
1845–1863. DOI: 10.1016/j.automatica.2004.05.015.
[7] T. Çimen: Systematic and effective design of nonlinear feedback controllers
via the state-dependent Riccati equation (SDRE) method. Annual Reviews
in Control, 34(1), (2010), 32–51. DOI: 10.1016/j.arcontrol.2010.03.001.
[8] J.R. Cloutier, C.N. D’Souza, and C.P. Mracek: Nonlinear regulation and
nonlinear H∞ control via the state-dependent Riccati equation technique:
Part 1, Theory; Part 2, Examples. In: Proceedings of the First International
Conference on Nonlinear Problems in Aviation and Aerospace, Daytona
Beach, FL, (1996), 117–141.
[9] R. Czyba and L. Stajer: Dynamic contraction method approach to digital
longitudinal aircraft flight controller design. Archives of Control Sciences,
29(1), (2019), 97–109. DOI: 10.24425/acs.2019.127525.
[10] E.B. Erdem and A.G. Alleyne: Globally stabilizing second-order nonlinear
systems by SDRE control. In: Proceedings of the American Control Con-
ference, San Diego, CA. USA, (1999). DOI: 10.1109/ACC.1999.786502.
[11] E.B. Erdem and A.G. Alleyne: Design of a class of nonlinear con-
trollers via state dependent Riccati equations. IEEE Transactions on Control
Systems Technology, 12(1), (2004), 133–137. DOI: 10.1109/TCST.2003.
819588.
[12] J. Geng, Y. Sheng, and X. Liu: Finite-time sliding mode attitude control
for a reentry vehicle with blended aerodynamic surfaces and a reaction
control system. Chinese Journal of Aeronautics, 27(4), (2014), 964–976.
DOI: 10.1016/j.cja.2014.03.013.
[13] K.D. Hammett, C.D. Hall, and D.B. Ridgely: Controllability issues in
nonlinear state-dependent Riccati equation control. Journal of Guidance,
Control and Dynamics, 21(54), (1998), 767–773. DOI: 10.2514/2.4304.
[14] A. Heydari and S.N. Balakrishnan: Closed-form solution to finite-
horizon suboptimal control of nonlinear systems. International Journal
of Robust Nonlinear Control, 25(15), (2015), 2687–2704. DOI: 10.1002/
rnc.3222.
[15] M.H. Korayem and S.R. Nekoo: Finite-time state-dependent Riccati equa-
tion for time-varying nonaffine systems: Rigid and flexible joint manip-
 FINITE-TIME SDRE CONTROL OF F16 AIRCRAFT DYNAMICS 575

ulator control. ISA Transactions, 54, (2015), 125–144. DOI: 10.1016/

j.isatra.2014.06.006.
[16] Q.M. Lam, M. Xin, and J.R. Cloutier: SDRE control stability criteria and
convergence issues: where are we today addressing practitioners’ concerns?,
In: AIAA papers, (2012), 2012–2475. DOI: 10.2514/6.2012-2475.
[17] Y.W. Liang and L.G. Lin: On factorization of the nonlinear drift term for
SDRE approach. In: Proceedings of the 18th World congress IFAC., 44(1),
(2011), 9607–9612. DOI: 10.3182/20110828-6-IT-1002.00808.
[18] K. Lu and Y. Xia: Finite-time fault-tolerant control for rigid spacecraft with
actuator saturations. IET Control Theory and Applications, 7(11), (2013),
1529–1539. DOI: 10.1049/iet-cta.2012.1031.
[19] D.G. Mitchell, R.H. Hoh, B.L. Aponso, and D.H. Klyde: The mea-
surement and prediction of pilot-in-the-loop oscillations. In: AIAA Papers,
(1994), 94–3670. DOI: 10.2514/6.1994-3670.
[20] C.P. Mracek and J.R. Cloutier: Control designs for the nonlinear bench-
mark problem via the state-dependent Riccati equation method. Interna-
tional Journal of Robust Nonlinear Control, 8(4–5), (1998), 401–433. DOI:
10.1002/(SICI)1099-1239(19980415/30)8:4/5.
[21] J.D. Pearson: Approximation methods in optimal control. Journal of
Electronics and Control, 13, (1962), 453–469. DOI: 10.1080/0020721620
8937454.
[22] E. Promtun and S. Seshagiri: Sliding mode control of pitch-rate of an F-16
aircraft. IFAC Proceedings Volumes (IFAC Papers-OnLine), 41(2), (2008),
1099–1104. DOI: 10.3182/20080706-5-KR-1001.00190.
[23] W.C. Reigelsperger and S.S. Banda: Nonlinear simulation of a modified
F-16 with full-envelope control laws. Control Engineering Practice, 6(3),
(1998) 309–320. DOI: 10.1016/S0967-0661(98)00024-0.
[24] J.S. Shamma and J.R. Cloutier: Existence of SDRE stabilizing feedback.
IEEE Transactions on Automatic Control, 48(3), (2003), 513–517. DOI:
10.1109/TAC.2002.808473.
[25] M. Siwczyński and K. Hawron: Optimum tasks and solutions for en-
ergy transmission from the source to the receiver. Bulletin of the Polish
Academy of Sciences. Technical Sciences, 66(5), (2018), 655–663. DOI:
10.24425/124281.
[26] M. Sznaier, J. Cloutier, R. Hull, D. Jacques, and C. Mracek: Receding
horizon control Lyapunov function approach to suboptimal regulation of
576 M. CHODNICKI, P. PIETRUSZEWSKI, M. WESOŁOWSKI, S. STĘPIEŃ

nonlinear systems. Journal of Guidance, Control and Dynamics, 23(3),

(2000), 399–405. DOI: 10.2514/2.4571.
[27] I. Tanyer, E. Tatlicioglu, and E. Zergeroglu: Output tracking control
of an aircraft subject to additive state dependent disturbance: an optimal
control approach. Archives of Control Sciences, 31(2), (2021), 267–286.
DOI: 10.24425/acs.2021.137418.
[28] K.K. Varanasi and S. Nayfeh: The dynamics of lead-screw drivers: low
order modeling and experiments. Transactions of ASME, 126, (2004), 388–
396. DOI: 10.1115/1.1771690.
[29] N. Wahid, M.F.A. Rahmat, and K. Jusoff: Comparative assesment using
LQR and fuzzy logic controller for a pitch control system. European Journal
of Scientific Research, 42(2), (2010), 184–194.
[30] N. Wang, J. Yu and W. Lin: Positioning control of linear actuator with
nonlinear friction and input saturation using output feedback control. Com-
plexity, 21(S2), (2016), 191–200. DOI: 10.1002/cplx.21797.
[31] F. Wang, Q. Zong, and B. Tian: Adaptive backstepping finite time attitude
control of reentry RLV with input constraint. Mathematical Problems in
Engineering, 5, (2014), 1–19. DOI: 10.1155/2014/801747.
[32] A. Wernli and G. Cook: Suboptimal control for the nonlinear quadratic
regulator problem. Automatica, 11(1), (1975), 75–84. DOI: 10.1016/0005-
1098(75)90010-2.
[33] Y.W. Liang and L.G. Lin: Analysis of SDC matrices for successfully imple-
menting the SDRE scheme. Automatica, 49(10), (2013), 3120–3124. DOI:
10.1016/j.automatica.2013.07.026.
[34] W. Durham: Aircraft Flight Dynamics and Control. First Edition, John
Wiley & Sons Ltd., 2013.
[35] M.V. Cook: Flight Dynamics Principles. Second Edition, Elsevier Aero-
space Engineering Series, 2007.
[36] Flight Control Design – Best Practices, RTO TECHNICAL REPORT 29,
RTO/NATO 2000 [Online]. Available: https://www.sto.nato.int>RTO- TR-
029 [Accessed: 6. July 2021]
[37] The F-16 combines interoperability and advanced technologies into a
proven design [Online]. Available: https://www.lockheedmartin.com/en-us/
products/f-16.html [Accessed: 6. July 2021].